JP2009205687A - Multiple device management method and system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a multiple device management method and a system that executes commands on one controller computer such as in a data center, to initiate operation on controlled computers containing agent software. <P>SOLUTION: A controller 202 communicates with nodes using a message format, and provides a schema to represent available nodes in the data center, their organizations into sets, and results of ongoing and completed operations. The schema is used to manage jobs to be performed on nodes and sets of nodes. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a method and system for managing a plurality of computer devices arranged on a network.

  This application claims priority to US provisional patent application filed June 11, 2001, application number 60/297473.

  Data centers provide computer resources for many types of organizations, including enterprises and various Internet-related service providers such as storage service providers, hosting service providers, and application service providers. A typical data center includes hundreds to thousands of computers that play different roles for different purposes.

  Managing a large number of computers can be expensive, time consuming, and error prone. For example, many service providers recognize that operating a data center is a labor intensive business because many routines are performed manually. As an example, bringing a new server online for a new customer can print a new customer order, install an operating system, install any application, configure the operating system and applications, It consists of letting the technician move to the data center and connect to the network and power supply. Thus, bringing a new server online is a long and expensive process and is prone to errors.

  A similar manual process is used when “resetting” an existing server configuration for a new customer via software. In fact, the cost of resetting is so high that some service providers consider it cheaper not to do so. A portion of the cost is spent in locating such unused systems through expensive manual data center audits. Similarly, rather than redeploying a computer that is no longer needed, leave the server computer in the data center (operating and consuming power) or attempting to redeploy the computer. It may be cheaper to completely dismantle.

  In short, operating a data center requires some compromises that are necessary for practical purposes but are less desirable. For example, instead of redeploying a computer, it may be cheaper to do so, which means that the data center is operating the computer system even when the computer system is no longer needed (While using power, air conditioning, and network ports).

  As another example, although expensive, the manual configuration of various servers is still a data center approach. However, reducing such costs through automation is an important effort that has not been so successful. This is unsuccessful, especially because such attempts have the difficulty of requiring the integration of multiple external products.

  Briefly stated, the present invention provides a method and system for multiple device management, which, among other things, executes a single command on one controller computer and includes one or more called nodes. Provides a mechanism for initiating operations on other controlled computers. An operation may include execution of a scripted set of commands, execution of a binary program, or some other type of operation. The mechanism works with a few sets of computers as if they were a single computer, which greatly simplifies the management of computer devices, for example, in a data center. Management costs are significantly reduced.

  In one embodiment, the present invention provides an architecture that includes a controller (eg, a process, etc.) on one computer that manages multiple other computers, each of the multiple other computers being manageable by the controller. Agent software to be included. In general, a controller provides a central representation of multiple nodes it manages and processes from this central representation for selected nodes individually or for each set to which a node can be assigned. Can start. The controller communicates with the node using a message exchange format, such as a format derived from XML (extensible Markup Language), using the underlying replaceable transport layer for network communication.

  The controller provides a prescribed way of representing available nodes in the data center, the organization of available nodes into sets, and the results of ongoing and completed operations. For example, continue to represent a node and a set of nodes (eg, grouped by an administrator, usually according to some criteria, such as administrative convenience, operational purpose, or other criteria) using a schema . Also, a record of the results of each process can be stored on a storage device accessible to the controller, along with the operations and jobs that are waiting and the operations and jobs that have been executed using the schema.

  Other advantages will become apparent from the following detailed description considered in conjunction with the drawings.

  In accordance with the present invention, a method and system for managing multiple devices is provided, which performs, among other things, a single command on one controller computer and one or more other controls called nodes. Provides a mechanism for initiating operations on the computer to be executed. An operation may include execution of a scripted set of commands, execution of a binary program, or some other type of operation. The mechanism works with a few sets of computers as if they were a single computer, which greatly simplifies the management of computer devices, for example, in a data center. Management costs are significantly reduced.

FIG. 6 is a block diagram representing a computer system in which the present invention can be incorporated. 2 is a block diagram illustrating a controller computer connected to a network for managing a plurality of nodes according to an aspect of the present invention. FIG. FIG. 6 is a block diagram representing various example components within a controller computer and various example components within one of the nodes managed by the controller computer in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. FIG. 4 represents a suitable defined schema for continuing the representation of available nodes, sets, operations, and / or jobs, etc., in accordance with aspects of the present invention. 6 is a flow diagram illustrating general logic for performing operations on one or more selected nodes in accordance with aspects of the present invention. 6 is a flow diagram illustrating general logic for performing requested processing on a node according to an aspect of the present invention. 3 is a flow diagram generally representing a result of an operation obtained, as well as a user interface, script, or process that initiated the operation on the controller that determines the result of the operation, in accordance with aspects of the present invention.

Exemplary Operating Environment FIG. 1 illustrates an example of a suitable computer system environment 100 on which the invention may be implemented. The computer system environment 100 is only one example of a suitable computer environment and is not intended to suggest any limitation as to the scope of use or functionality of the invention. Neither should the computer environment 100 be interpreted as having any dependency or requirement relating to any one or combination of components illustrated in the exemplary operating environment 100.

  The invention is operational with numerous other general purpose or special purpose computer system environments or configurations. Examples of well-known computer systems, computer environments, and / or computer configurations that may be suitable for use with the present invention include personal computers, server computers, handheld devices or laptop devices, tablet devices, multiple devices Processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, routers, distributed computing environments including any of the above systems or devices, etc. It is not limited to them.

  The invention can be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The invention is generally intended to be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be in local and / or remote computer storage media including memory storage devices.

  Referring to FIG. 1, an exemplary system for implementing the present invention includes a general purpose computing device in the form of a computer 110, which operates as the controller computer 210 of FIG. 2 for controlling nodes. Can do. The components of the computer 110 can include, but are not limited to, a processing device 120, a system memory 130, and a system bus 121 that couples various system components including the system memory to the processing device 120. Not. The system bus 121 can be any of several types of bus structures including a memory bus or memory controller, a peripheral bus, and a local bus using any of a variety of bus architectures. By way of example, and not limitation, such architectures include the Industry Standard Architecture (ISA) bus, the Micro Channel Architecture (MCA) bus, the Enhanced ISA (EISA) bus, the Video Electronics Standards AES eZ and the EZeZ A Peripheral Component Interconnect (PCI) bus, also known as a bus, is included.

  Computer 110 typically includes a variety of computer readable media. Computer readable media can be any available media that can be accessed by computer 110 and includes both volatile and nonvolatile media, removable and non-removable media. . By way of example, and not limitation, computer readable media can include computer storage media and communication media. Computer storage media is volatile and non-volatile, removable and non-removable implemented in any method or technique for storing information such as computer-readable instructions, data structures, program modules, or other data. Including both media. Computer storage media can be RAM, ROM, EEPROM, flash memory or other memory technology, CD-ROM, digital versatile disk (DVD) or other optical disk storage, magnetic cassette, magnetic tape, magnetic disk storage or other magnetic storage device , Or any other medium that can be used to store the desired information and that can be accessed by the computer 110. Communication media typically embodies computer readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. The term “modulated data signal” means a signal that has one or more of its characteristics set or changed in such a manner as to encode information in the signal. By way of example, and not limitation, communication media includes wired media such as a wired network or direct-wired connection, and wireless media such as acoustic, RF, infrared and other wireless media. Any combination of the above is also included within the scope of computer-readable media.

  The system memory 130 includes computer storage in the form of volatile and / or nonvolatile memory such as read only memory (ROM) 131 and random access memory (RAM) 132. A basic input / output system 133 (BIOS) that contains basic routines that help to transfer information between elements within the computer 110, such as during startup, is typically stored in the ROM 131. The RAM 132 typically includes data and / or program modules that are immediately accessible to the processing unit 120 and / or that are currently being operated upon by the processing unit 120. By way of example and not limitation, FIG. 1 shows an operating system 134, application programs 135, other program modules 136, and program data 137.

  The computer 110 may also include other removable / non-removable volatile / nonvolatile computer storage media. By way of example only, FIG. 1 illustrates a hard disk drive 141 that reads or writes to a non-removable non-volatile magnetic medium, a magnetic disk drive 151 that reads and writes to a removable non-volatile magnetic disk 152, and Illustrated is an optical disk drive 155 that reads from and writes to a removable non-volatile optical disk 156, such as a CD ROM or other optical media. Other removable / non-removable volatile / non-volatile computer storage devices that can be used in exemplary operating environments include magnetic tape cassettes, flash memory cards, digital versatile disks, digital video tapes, solid state Examples include, but are not limited to, RAM, solid state ROM, and the like. The hard disk drive 141 is normally connected to the system bus 121 via a non-removable memory interface such as the interface 140, and the magnetic disk drive 151 and the optical disk drive 155 are usually systems with a removable memory interface such as the interface 150. Connected to the bus 121.

  The drive described above and shown in FIG. 1 and associated computer storage media provide computer 110 with storage of computer readable instructions, data structures, program modules, and other data. In FIG. 1, hard disk drive 141 is shown as storage operating system 144, application program 145, other program modules 146, and program data 147. These components can be the same as or different from the operating system 134, application program 135, other program modules 136, and program data 137. Operating system 144, application program 145, other program modules 146, and program data 147 are numbered differently in this specification to indicate at least different copies. A user may enter commands and information into the computer 20 through input devices such as a tablet, commonly referred to as a mouse, trackball or touch pad or electronic digitizer 164, microphone 163, keyboard 162 and pointing device 161. . Other input devices not shown in FIG. 1 may include joysticks, gamepads, satellite dishes, scanners, and the like. These and other inputs are often connected to the processing unit 120 via a user input interface 160 coupled to the system bus, but other interfaces such as a parallel port, game port, or universal serial bus (USB) and It is also possible to connect with a bus structure. A monitor 191 or other type of display device is also connected to the system bus 121 via an interface, such as a video interface 190. The monitor 191 can be integrated with a touch screen panel or the like. The monitor and / or touch screen panel can be physically coupled to a housing in which the computing device 110 is incorporated, such as in a tablet personal computer. Further, a computer such as the computer device 110 may include other peripheral output devices such as a speaker 195 and a printer 196 that can be connected via an output peripheral interface 194 or the like.

Computer 110 typically operates in a networked environment to control a number of remote server computers 180, referred to as nodes 204 or nodes 204 ₁ -204 _n (FIGS. 2 and 3). Such nodes typically include some or all of the components of computer 110 described above. A node may not include user interface mechanisms such as a keyboard, mouse, display. Such nodes can execute various programs that typically provide services to other computers or users in a networked environment. Examples of such programs include programs that provide web pages or manage databases. By way of example and not limitation, FIG. 1 shows that a remote application program 185 is resident on the memory device 181. It will be appreciated that the network connections shown are exemplary and other means of establishing a communications link between the computers can be used.

Multiple Device Management FIG. 2 shows a controller computer 202 (which may correspond to the computer system 110 of FIG. 1) and a plurality of managed computer devices called nodes 204 ₁ -204 _n (remote computer 180 of FIG. 1). FIG. 1 is a block diagram generally representing an architecture 200 including: However, aspects of the present invention are suitable for use in a data center, but such aspects are applicable in any environment where one or more tasks need to be performed on one or more computers. Please note that. For example, a user with multiple computers and services (such as PDAs, mobile computers, desktop systems, home media center systems, etc.) can use the present invention to perform operations across such multiple devices. . Thus, although generally described in a data center environment, the present invention should not be limited to any particular configuration or configurations, but substantially any configuration of the computer devices presented above. As well as personal computers, servers, routers, and various storage devices.

A typical data center includes a number of different computers. In order to make management of the data center easier, an administrator or the like designates one of the computers as the controller computer 202 (or the controller 202). In particular, in accordance with one aspect of the present invention, the controller computer 202 is a master of the computers in the data center, how the computers are configured, and what operations can be performed on the computers. Hold records. The controller can store this data in a locally connected data store (eg, as generally represented in FIG. 3) or on a remote data store. Node 204 ₁ to 204 _n includes a computer in a data center can be managed by the controller 202. The nodes include node programs 206 ₁ -206 _n that perform the actual work of the data center, such as providing web pages or managing databases.

In accordance with one aspect of the present invention, an administrator applies a logical organization to computers to operate hundreds or thousands of computers in a data center. That is, the administrator groups the various nodes 204 ₁ -204 _n into sets that represent the nodes as a logical structure instead of just a list of all the nodes in the data center. Multiple such structures of computers for a single data center can exist, for example grouped by customer or function, and nodes can belong to multiple sets. The set information is stored by the controller service 208, etc., in the data store 302 (FIG. 3) maintained by the controller, for example, by placing the computer in a set named with Unicode characters. Through the set, an administrator can operate a logically organized group of computers as if they were a single computer. Note that if there are multiple controllers in the data center, grouping the controllers can also be implemented.

  To establish and maintain a set, such as a management program 210 including one or more applications, processes, threads, objects, and / or other software components, etc., via one or more user interfaces, etc. , Allowing the administrator to interact with the controller service 208. Management program 210 allows an administrator to create a set, add nodes to the set, remove nodes from the set, delete the set, list the current node in the set, and list the set to which the node belongs Will be able to.

  For example, data center management corresponds to various conceptual structures in the configuration of node computers, the physical layout in which the node computers are identified by their physical location within the data center (eg, cages, racks, and slot locations). A structure such as a configuration can be applied. This view of the data center allows the management department to find the physical location of the node, for example, when it is necessary to replace the node. Alternatively, the physical layout set allows the management department to specify the installation location of a newly purchased computer (node).

  There are other ways to organize the nodes in the data center, including a logical structure that organizes the nodes for each customer that uses each node. This allows an administrator to map between a customer and one or more nodes currently used by that customer to track or monitor the use of a particular customer, or all nodes of that customer You will be able to make changes. Another logical view of the data center is to organize nodes by function, which is an application that runs on the node. For example, certain nodes can be grouped because they are web servers, media servers, or databases. By looking at the data center for each function, the management department can perform a certain process (such as applying a patch to a specific application) only for the nodes that need it. Note that the above is just an example. For example by network topology (nodes connected to a specific switch or load balancing), by VLAN, by customer service level, by machine function and / or by installed OS and service patch level. As such, there are various other ways in which a data center management department may wish to organize nodes in its own data center. Note that any given node can be grouped into any number of sets, as presented in the following example table.

  Consistent with this aspect of the invention, the ability to group nodes into sets (eg, distinguished by simple names) simplifies the performance of various operations. For example, the set can identify the node computers that are in use by a customer when that customer reports a problem, thus quickly accessing the status of those node computers and solving the problem. it can. Monitoring information collected from nodes is aggregated into a form that can be viewed via a website for a specific customer so that the customer can access and view the status of their own machine Is possible. The node can monitor for resource exhaustion (eg, running out of disk space) and use the machine to generate alerts for customers. Alerts are reported to the customer (eg, via email) and / or can be used by sales staff to contact the customer and encourage them to upgrade the configuration to better hardware or additional servers, for example. is there.

  In accordance with another aspect of the invention, for managing one or more nodes in the data center by initiating operations on the controller server 208, such as via an initiation process 304, to manage the data center. Operation is performed. In FIG. 3, the start process 304 is represented as part of the management program 210, but it will be appreciated that the start process and the management program can be separate, in fact, the start process can be a script or other Note that the executable code can be Further, for simplicity, FIG. 3 shows only one node 204, but it is noted that the controller 202 can control multiple nodes, for example, at one time or in some order, as will be appreciated. I want.

  As generally represented in FIG. 3, on the controller computer 202, the controller program 306 of the controller service 204 of the present invention is accessed and executed via the schema interface 310. This schema (shown generally in FIGS. 4-12) provides a representation of the available nodes, sets, available operations in the data center, and the results of performing each operation. In general, on the controller computer 202, the controller service 208 can be considered to include a schema interface 310, a controller program 306, and a data store 302. Any process, script, or user interface that wishes to access the controller service 208 does so via the schema interface 310. The controller program 306 determines processing to be performed. This process includes accessing the data store 302 and / or communicating with one or more nodes (such as node 204) via the transport layer 212. For example, the result data and other information needed to provide the administrator with representations are generally maintained in the data store 302. The data store can be accessed by an administrator via the schema interface 310. In a typical embodiment, the schema interface is a Common Information Model (CIM) for representing computer processes and computer devices, such as described in US patent application Ser. No. 09/020146, assigned to the assignee of the present invention. It is possible to derive from the standard.

  In general, to perform an operation, first, one or more nodes or sets of nodes on which the operation is to be performed are selected, such as via the management program 210. The selection can specify a single computer, all computers in a set, all computers in several sets, or a mixture of individual computers and computers in a set. From the selection, for example, if the same node is specified through two or more selected sets, any duplication can be removed. Whereas a set is a representation selected by an administrator of a data center organization, a selection is generally a collection of computers that are temporarily designated to have at least one operation performed thereon. For this reason, it should be noted that the selection is different from the set. In other words, set members (and the set itself) can change over time, but are a continuing representation of the organization of computer nodes, whereas selections are intended to meet the intentional goals of a particular operation. A temporary set of computer nodes to represent. A selection is made when an operation or series of operations needs to be performed on a given collected number of computers. Upon completion of the operation or series of operations, the representation of the selection can be deleted without losing information regarding the organization of the computer in the data center. The selection can be considered a “temporary” set. In some embodiments, the selection can be limited to degenerate cases where the selection can include only a single set or a single node.

  Specific operations that can be performed by the administrator (eg, via management program 210) include creating a selection, deleting the selection, and listing the computers being selected. In one embodiment, the selections are stored on the controller 202 and each selection is identified by a unique serial number assigned when the selection was made. The selection can support operations in a data center that includes thousands of computers, and each selection can include any number of computers, from one to the total number of members in the data center. Records are kept to keep track of which operations are performed on which devices, thereby providing an audit of what has been performed. This audit is usually archived and purged periodically.

  Note that in a data center environment, the controller 202 is the control point of the data center (at least a portion thereof) and is a repository for information about the data center. As a result, in such an environment, the controller needs to be highly usable and recoverable in the event of a disaster. With regard to availability, the controller can include a shared disk cluster environment of multiple clustered machines, which can affect the operation of the controller 202 even if several clustered computers fail. It is possible not to give. With regard to disaster recovery, the state of the controller can be backed up, including details of the results of the group, script, script code, and previously completed jobs and jobs being processed. The backup process can be documented and encapsulated in a script that an administrator or other service can use to perform regular backups. After the backup, the controller state needs to be restored on the new controller, and furthermore, the new controller needs to manage nodes that were previously controlled by the previous controller. At the same time, security is provided to prevent unauthorized controllers from keeping nodes under management.

  Scripts that can be executed on the node are stored on the controller 202 (eg, in a script database in the data store 302 or elsewhere). Scripts can be written to use any script host that is available on the target node. The script database contains information about the scripts available for execution on the remote node, and the scripts themselves can be stored on the controller's file system, and the database contains the path to the scripts. In one embodiment, the script is placed anywhere on the controller's file system. Creation and editing of scripts on the file system, creation of script entries in the script database, editing of script entries, deletion of script entries, jobs that may use a script (described below) ) Creation, job deletion, job editing, job execution, job status and result search.

In accordance with one aspect of the invention, to perform an operation or series of operations on one or more selected nodes, once a selection is made, the controller service 208 may receive a message (or a job containing a job to be executed) (or Multiple messages). The message is preferably formatted using XML, message exchange protocol and transport layer protocol 212 on the controller, and corresponding protocol 214 ₁ on each of the one or more nodes specified in the selection. ˜214 _n are used to transmit to each selected target node (eg, node 204 in FIG. 3). For simplicity, the controller program 306 may be considered to communicate directly with the agent service 218 including the agent program 312 on the node computer 204 using the defined XML message exchange protocol 314 (FIG. 3). Meanwhile, the transport layer 212 on the controller 202 can be considered to communicate directly with the transport layer 214 on the node 204 using the transport protocol 316 (FIG. 3). The transport layer 212 includes any system that can reliably send messages from the controller to the plurality of nodes 204 ₁ -204 _n and that the node 204 can use to send messages to the controller 202. . For example, multicast can be used to send messages from the controller to multiple nodes. Unicast can be used if multicast is not available or if the target is a single node. An alternative transport layer is to use the standard HTTP protocol to include messages and use unicast.

In one embodiment, communication between the controller 202 and the node 204 sends a script to the agent service 218 using the Multiple Device Management (MDM) protocol, returns the result of executing the script, and returns a binary executable code. Tell agent service 218 to execute, perform other operations on node 204 via agent service 218, and send alarm and event information from agent service 218 to controller 202. In this embodiment, the MDM protocol operates on a transport layer protocol that reliably provides network communication between the controller and one or more agents, such as by using Sender-side Guaranted Multicast (SGM). . Although the present invention is primarily described by logical communication via message exchange and / or transport protocols, the underlying physical network 216 (whether wired or wireless) connects the controller 202 and nodes 204 ₁ -204 _n . Note that it is understood to connect.

  As an example of one general type of operation, an administrator executes a set of commands (eg, a script) on one or more selected nodes via the controller computer 202. This is generally achieved by selecting a node and sending a message to the selected node. In addition to scripts, the message exchange protocol 314 can be extended to allow alternative types of operations such as the execution of precompiled programs, or some other operation.

  In accordance with another aspect of the invention, node 204 includes an agent service 218 that includes an agent program 312. In general, the agent service 218 is responsible for processing upon request of the controller 202 and for sending alarms and events to the controller 202. For this purpose, the agent program 312 receives a communication (message) from the controller 202 and determines how to perform one or more requested processes specified in the message (if any). If necessary, the process is passed to an appropriate execution engine 320 together with an arbitrary argument (parameter). As an example, when the controller 202 sends a script to the agent program 318, the execution engine 320 is typically a corresponding script interpreter. The agent program also allows some other operation that is part of the protocol between the controller and the computer, such as the execution of binary executable code on the node 204. Instead of sending a binary executable code, the controller 202 preferably sends a network address for the code for the node 204 to execute the code from its location or download the code. Note that it performs. For completeness, node operating system 220 as well as controller operating system 222 are shown.

  In addition to script and program execution, other operations such as reboot, shutdown, or suspend (go to a low power state) may be required of the controlled node. This is because if such an operation is performed immediately (such as via a script), the node 204 may not be able to provide a result to the controller. Instead, such operations are performed by a special function handling component 322 that first communicates to the controller 202 a result that basically states that the command message has been received and understood. After sending the result, the node performs an appropriate process to satisfy the request.

  In short, the MDM protocol allows the controller 202 to perform certain operations such as executing a script on the agent, using a standard scripting host, or executing binary code on the agent. One or more nodes can be requested. The protocol does not understand the operations to be performed, which keeps the protocol simple and avoids special case operations. However, some operations have an impact on the operation of the protocol itself and are therefore expressed in the protocol. Such special operations may include rebooting, suspending, or shutting down the node computer. The MDM protocol also allows the controller to exchange management information, such as when it first communicates with the node to control the node.

  The MDM protocol is also used to send back to the controller 202 the results of executing scripts or binaries or special operations. Node alerts and node events can be sent via the MDM protocol and heartbeat to periodically inform the controller that the node is operating properly. The interval information is configurable and the controller 202 can send the interval information to the node via the MDM protocol.

Upon receiving the result of the operation, the controller 202 holds a record of the operation. For this purpose, as coming result entered from node 204 ₁ to 204 _n, the result is extracted from the returned messages, and stored in the database 302. This provides a complete record of each operation ongoing for each node. An administrator or process (such as a script) can query this information on the controller 202 to determine the success of the operation on each node and, if necessary, investigate or resolve the failure. The message format also provides a way for the controller to maintain information about the state of each node, including the status of any alarms on the node.

  As will be appreciated, various agent service operations, including the execution of scripts and other processes, give the administrator considerable flexibility to perform any operation on the servers in the data center. For example, many normal management operations provide pre-built, tested scripts with a controller or node computer that can be a preconfigured server. Such scripts provide data center managers with the ability to manage servers in the data center for general operation and are provided with the controller 202 ready for use. Additional scripts are provided with specific device servers and the like (including new types of devices such as cache devices). In such a case, the administrator need only load the script into the controller 202 and configure the controller program 306 to know about this new script. Further, for example, the node computer can be shipped with a related script mounted on the basis of the function, and the script can be automatically provided to the controller 202. For special situations, custom scripts can be written and tested and added to the controller 202 by configuring the controller to know about the new script.

  As described above, each node computer includes an agent service 218. As with any given controller, agent installation is provided with the node, eg, a computer can be shipped as equipment that does not require any installation. Alternatively, the agent service can include separate software that needs to be installed, such as via a supplied installation script that installs on each computer to be managed.

  Provided by Universal Plug-n-Play (uPnP, the standard for automatic detection computer devices on the network) when the computer including the agent service 218 is booted to simplify the processing burden on the data center It is possible to automatically broadcast its presence over the network via the detection component 330, such as via an automatic detection protocol. On nodes with multiple network interface cards (NICs), the NICs used to perform broadcast communication can be limited, for example, when multiple NICs are used, until the next reboot. Only the first NIC containing the responding controller is used.

  The controller 202 recognizes the broadcast via the detection listening process 332, and if the computer is not already in the controller's representation of the computer that the controller knows, the computer that will broadcast is added. The controller refers to the computer as a node. For this purpose, when the controller 202 recognizes the auto-detect broadcast, it determines whether it already knows about the node. If you already know, that node has been booted before in the data center and is in the uncontrolled node or list of controlled nodes. If it is a controlled node, the controller 202 reestablishes control of that node. In either case, the controller marks the node record to indicate that the node has been booted on the network. If the node is not known by the controller, the controller 202 responds by adding information about the node to its internal database 302. The information acquired by the controller 202 is a unique identifier of the node (such as a globally unique hardware identifier such as a BIOS GUID or MAC address of the network interface card). When a node is newly detected, it is first considered an uncontrolled node and the administrator can make a decision whether to make that node a node controlled by the controller 202.

  In addition to simplifying control, auto-discovery further provides a way to use new machine auto-configuration. For example, an administrator can either inventory a new set of commands, for example a new system, apply the currently required host fixes for all new machines, or You can specify whether to run commands that configure the machine for different customers. The administrator can then write a script that consumes the events that occurred when a new computer was added on the controller and took appropriate action. After discovery, the administrator can list all new nodes in the data center and can add them to the set or perform operations on those nodes as needed. As you can easily understand, you have to manually enter information about all new computers (eg name, serial information and IP information) by using an auto-discovery mechanism to find new computers in the data center. This avoids possible data entry errors. In this way, the administrator maintains an online reference list of available computer nodes that helps to maintain inventory information regarding data center content.

  In order to provide security, the present invention protects against controller software and agent software ("Rogue Controller" and "Rogue Agent", respectively) that are not allowed by the data center administrator. Controllers and agents under the control of the data center administrator are referred to as “approved controllers” and “approved agents”, respectively. For this purpose, nodes that broadcast auto-discovery information on a network where there may be several controllers are configured to accept only responses from the first responding controller. When a node receives a response using public and private key technology, it will only accept control information from that controller and other authorized controllers. Usually this will be only the controller that first controlled the node, but it is expected that the controller will fail and need to be replaced. However, such a replacement controller is authorized via a secret key. The first controller to respond may be a rogue controller, in which case the node will be lost from data center control, which is appropriate on top of the agent before the agent is first booted. Note that this can be avoided by deploying secure certificate data. If a new node is taken by a rogue controller, there may be a loss of resources for the data center administrator, but the rogue controller has no access to customer information on the node.

  Once authorization is established between the controller and the node, the node becomes controlled and the authorized controller can take over control of the node (eg to replace the original controller). However, the unauthorized controller cannot take over control of the node. An unauthorized computer can also sniff the network, so sensitive information is not passed over the network without being encrypted, and only the target node of the secured message encrypts the message. It can be released. The operation can specify whether the communication needs to be encrypted when invoked. This can be done at the request of the initiating process or because the script is marked as requiring encryption in the script database.

  The data store 302 held by the controller 202 is configured as a defined schema. In general, the controller maintains details of nodes in the data center, set information, available operations, current, ongoing operations, and results of completed operations. Although not required, for convenience, this database schema is configured similarly to the object model schema through which management program 304 interfaces with controller program 306 and data store 302 (via schema interface 310). . In one embodiment, the object model schema represents information in the form of defined objects, but in general, the database schema holds records etc. used to construct objects when queried. Please note that. The schema is generally represented in FIGS. 4-12 and is further described in Appendix A. However, the schema described is only an example, and the present invention does not cover any particular way of holding or presenting data.

  Thus, the controller 202 maintains a certain amount of information about each node in the database. For example, each node is represented by a device data structure including communication-related data (for example, TCP-IP host name), a unique identifier, and various other data. Most of this information is a cached representation of the state of the node, such as the name of the node. The cached information may include items to present in the following table, which details how the information is updated on the controller when the information changes on the node Including.

  Other items of information about nodes held on the controller include:

  A node is identified on the controller by various information such as the node name (short form of the node name, for example, “server07”) and the MAC address of the NIC card on the node. The node name can be used as a unique identifier for the node record on the controller. Each time the node boots, both of these information are sent from the node to the controller using auto-discovery. If auto-discovery is not available on the network, the administrator needs to manually add a node record that includes at least the node name.

  The node name can also be used by the controller 202 to communicate with the node. In one embodiment, the controller uses DNS to resolve names to IP addresses. This means that the administrator needs to ensure that the controller has access to the DNS server that maps the node name to the IP address on the management NIC. The DNS server may know the current server name and IP address based on static data or using dynamic DNS. Note that if the management network currently does not have a DNS server, the controller itself can be used as a DNS server. A similar process for dynamic DNS can be used, such as the process of watching for IP changes on a node and sending updates to the controller. When the node becomes controlled, the controller establishes a permanent connection to the node.

  Auto-discovery is based on a network between a node and a controller that supports forwarding multicast detection packets, for example, packets are sent by a node to a predefined (fixed) multicast IP address and port. Because not all data center environments support multicast forwarding (eg due to router limitations or policies), data centers should use a controller per multicast domain (usually one per subnet) Controller), or other modes, such as operating without automatic detection. In that case, some operations that are otherwise automatic are performed manually.

  When the managed node reboots, the controller breaks any permanent connection to the node. In the current embodiment, the controller 202 waits for an auto-detect packet from the node rather than attempting to automatically re-establish communication. When controller 202 receives this packet, it knows that the node is available and re-establishes a connection with that node. If auto-detection is not available for work, the administrator can use the appropriate method, eg, Controller. Manually re-establish communication between the controller and the rebooted node, such as via a RecoverManagedNode method. In general, methods do not return a value, so if desired, the caller uses language-specific methods to determine if the method (or other operation) has failed, as described in Appendix A. Note that there is a need to do.

  The object model that implements the schema includes three main features: sets, devices, and jobs. A set represents a group of devices. Each set has a unique name and may contain no devices or may contain one or more devices. A given device can be in multiple sets. A set is represented only on the controller, and in one embodiment, the device has no knowledge of which set it is in and has been added to or removed from the set When you do not receive notifications. The set is implemented as an object “Sets”. It should be noted that in one embodiment, only a device can be a member of a set, while in other embodiments, a set can be a member of another set.

  A device is an individual server in a data center that can be managed by a controller. In one current embodiment, the only devices that can be managed are computers with agent software installed. The device is identified by its name, and the controller uses DNS to resolve this name to an IP address to communicate with the device.

  Device information is stored in several objects. The Device object stores basic information including the name, while the MAC address is stored in a DeviceHWAddrs object linked to the Devices object. Other hardware information (such as SMBIOS GUID and disk drive signature) can also be stored in an instance of this object. In order to differentiate between different types of information, the DeviceTypes object holds definitions for different types of hardware addresses. Each instance of DeviceHWAddrs is linked to a corresponding instance of device type (DeviceTypes). An IP address associated with a specific MAC address can also be stored in an instance of DeviceHWIPAddrs associated with a DeviceHWAddrs instance that represents the specific MAC address. Currently, IP address information is not stored for MAC addresses.

  A job generally refers to two types of job information: a job template that is ready to be executed and a job history that is a previously executed job. The job template is stored in a job instance (JobInstances) object. An instance of this JobInstances object is identified by a combination of two properties, a job name and a job identifier. The job template is stored with a unique name in the Job Invocations with a zero job identifier and a zero job identifier. Job template properties (apart from job identifiers) can be edited by the user. In the current embodiment, no other object is used to store the job template.

  The job history is stored in several objects. The JobInformations object stores basic information including name, job identifier, and time the job was executed. The job identifier is a non-zero value. The name can be blank and can be the same as other JobInvocations instances (since the job identifier uniquely identifies the job history). Each JobInvocation representing a job history is linked to an instance of a Jobs object. This represents the status of the job regardless of the number of devices on which the job was executed. This can be considered the “parent” for the status of the job on each of the individual devices. The actual status for jobs on each device is stored in a further instance of one Jobs class per device. Each of these instances is linked to a parent Jobs instance and can be considered a “child” of the parent instance. This creates a two-level parent-child relationship that can be extended to further levels. Note that if a job is not run on any device (because the set on which the job is run is empty), the parent Jobs instance is not linked to any child Jobs instance.

  The Jobs instance corresponding to the job status on an individual device does not store the actual output of the job. Instead, the output is stored in one or more instances of a job log (JobLogs) object. Each instance of JobLogs stores a portion of the job output. By using the sequence property of this object to arrange the partial outputs in order, the complete output can be reconstructed. JobLogs stores three types of output including standard error output, standard output, and exit status when the job actually exits. A JobLogs instance is created when output from a job enters the controller. Related to a given (child) Job record because the job did not start on the device (in which case the Jobs record contains an error indication) or no output status or exit status has been received from the device yet Note that it is possible that no JobLogs exist.

  As can be seen from FIGS. 4-12, Appendix A and the above description, architecture, structure, and protocol of the present invention allow considerable flexibility and efficiency in managing a data center. A first example can be seen when installing an upgrade or patch on several computers. Instead of having the administrator go to each computer (possibly remotely) and execute the upgrade package, in the architecture and structure of the present invention, the administrator goes to the controller (possibly remote). In formula), the target node can be selected from the set and node list maintained on the controller and the upgrade can begin, and in the future this may be done at once. After the upgrade is initiated, the administrator examines the results on the controller to see which nodes have failed to complete the update (if any). This significantly reduces the effort required to perform the upgrade, automatically maintains an auditable log of the results of the upgrade, and (for example) an upgrade has previously been manual If a specific process that had to be input to each computer is required, the possibility of an error is also reduced. Further, aspects of the present invention can be used to store operations in a controller and test them on a test system before performing the same operations on a production system.

  A second example of this flexibility and efficiency can be seen when adding a new server for a new hosting customer. Instead of manually installing the operating system and then configuring the computer for the customer, in various aspects of the present invention, the administrator can turn on a number of power supplies that are turned on but not actually used. You can maintain any available “spare” computers. When a new customer signs up, a set of operations can be initiated (automatically or manually) to select one of the spare computers and configure it for a particular customer. This configuration also includes configuring any additional equipment (switches or load balancers) necessary to service the customer. Since the configuration steps are performed by scripts rather than manually, the risk of errors is significantly reduced. In addition, the configuration step can be performed automatically, such as automatically configuring the system for the customer after the customer completes the purchase request on the data center website (for example). Enables data center managers to configure the system.

  A third example can be seen when monitoring and collecting data regarding the status of a group of computers. By manually setting up the collection system, operations can be performed automatically on each of several computers rather than having to collect data from several computers. The operation can periodically report usage information to the controller, which stores the information in storage, such as a data store or other suitable memory. Information can be analyzed at a later point in time by the administrator to understand how the computer is operating, to investigate the reason for failure in the service, or for another suitable purpose.

  Turning now to the description of the operation of the present invention with particular reference to FIGS. 13-15, to perform operations on several nodes, as represented by step 500, a target node is first selected and the operations to be performed are selected. A containing job is created for that node. Selection may include selecting individual nodes, multiple individual nodes, sets, multiple sets, or any combination of the foregoing. The operations to be performed can be scripts, binary programs, or other types of tasks. The initiating process 304 (which can be an application, a web UI, a user via the command line, or an automated policy in a script or rule) makes a selection and selects an operation Please keep in mind. A job 404 is created in the data store for this operation on the selection.

  When run, the job is started with any provided arguments, as represented by step 504. When started, at step 506, the controller 202 creates a message that includes information about the task (eg, in the case of a script, the script itself and any parameters). This message is then associated on the controller 202 with the target node.

  The message and representation of the target node is used by the controller to initiate the operation, which in one embodiment creates an appropriate XML message at step 508 and targets the message using the transport layer at step 510. Reached by sending to the node. At this point, the transport layer 212 on the controller 202 is given information about the target node and a message to be sent to that node. Control is returned to the initiating process 304 (the user interface, script, or process that initiated the operation on the controller 202) so that the initiating process does not have to wait for the operation to complete on all nodes. Please keep in mind. This allows the initiation process 304 to initiate multiple operations as desired and collect the results of these operations later.

  The message is passed to the transport layer, which is responsible for ensuring that the message arrives at the correct node, eg via a TCP / IP network. When received at the target node, the job starts as represented by step 512.

  FIG. 14 generally represents a job executed on a node, such as node 204, starting at step 600 where a message is received. In FIG. 14, the left side of the vertical bar shows processing performed by the agent program or agent service on the node, while the right side shows processing performed by the operating system of the node. On the node, the agent's transport layer receives the message and passes the message to the agent service 218 for interpretation. In general, the agent service 218 determines how to perform the process, performs the process, and the result is returned to the controller via the transport layer.

  More particularly, step 602 represents extracting an operation from the message, and step 604 represents determining an execution engine to use. For example, different types of scripts require different execution engines and need to download binary code before execution. In any case, step 606 represents sending any arguments passed in the operation and message to the determined execution engine, while step 608 represents the operating system performing the operation.

  Step 610 represents the output of the operation collected and / or otherwise stored, the output being anything from a simple success or failure to the collection of data requested by the operation Is possible. Step 612 creates a message (eg, of the MDM protocol in XML format) and step 614 returns the message to the controller 202 via the transport layer. In this scheme, any result of the operation (such as output from a script or program) is formatted into a message and returned to the controller using the transport layer.

  FIG. 15 illustrates how an initiating process 304 that initiated an operation on the controller 202 can determine the result of the operation on each node. Generally, via steps 700-704, the controller maintains a record of operations. For example, as the result comes in from the node at step 700, the result is extracted from the message returned for each operation at step 702, and at step 704 the data store 302 (or other suitable storage on controller 202). ). This activity provides a complete record of the ongoing operation for each node.

  As represented through steps 708-712, an administrator or process (such as a script) that is not necessarily the initiating process 304 queries this information from the controller 202 to determine the success of the operation on each node and other data. And, if necessary, investigate or resolve the fault.

  As can be understood from the foregoing detailed description, a method and system for managing multiple devices that facilitates the management of computer devices within a data center or the like is provided. The method and system are very flexible, efficient, and significantly reduce the costs associated with managing multiple computing devices.

  While the invention is susceptible to various modifications and alternative constructions, certain illustrated embodiments thereof are shown in the drawings and have been described above in detail. However, it is not intended that the invention be limited to the particular forms disclosed, but on the contrary, the invention covers all modifications, alternative constructions, and equivalents falling within the spirit and scope of the invention. To do.

Appendix A
In one common MDM (Multiple Device Management) embodiment, a schema interface can be derived from a Common Information Model (CIM) standard for representing computer processes and computer devices. For example, such an embodiment on a Microsoft® Windows® operating system is referred to as a WMI schema. The various information herein describes such a suitable schema and other relevant data, but should be considered merely as examples.

Definition of identifier Set name The set name uniquely identifies the set. The set name is Sets. It is stored in the Name property and can be up to 256 Unicode characters. Any character is valid, but unprintable characters are not recommended. The name is case-insensitive, but the case is preserved. The set name must be unique among all set names.

Device Name The device name is used to uniquely identify the device and locate the device's IP address to communicate with the device. The device name is Devices. Stored in the Name property. The device name can be up to 256 Unicode characters. Any character is valid, but unprintable characters are not recommended. The name is case insensitive, but the case is preserved. The set name must be unique among all device names. Note that the device name and set name have the same number of characters because they are stored in a single field in the JobInformations object.

Job name The job name is used to uniquely identify the job template, and is used as one piece of identification information regarding the job history record. The job name is JobInformations. Name property and Jobs. Stored in the Name property. The job name can be up to 50 Unicode characters. Any character is valid, but unprintable characters are not recommended. The name is case insensitive, but the case is preserved. For job templates (jobs with a job identifier of zero), the job name needs to be unique. However, the same name can be used multiple times in the job history record depending on circumstances.

Device type name The device type name is used to identify the type of device. The device type name is Devices. Stored in the Type property, and the DeviceTypes. Defined in the value of the Name property. The device type name can be up to 50 Unicode characters. All characters are valid, but unprintable characters are not recommended. The name is case insensitive, but the case is preserved. The device type name must be unique among all instances of DeviceTypes.

Description The description is typically 256 Unicode characters (but see the individual object definitions for specific values). Any Unicode character is valid, including carriage return and newline characters.

Objects Each object can contain properties and methods, and can have associations with other objects. In one embodiment, objects, properties, methods, and associations are exposed via WMI. As described below, additional association classes are used to implement associations.

  Each object definition begins with a summary, followed by a definition of the form:

The conditions under which a class can be created and the conditions under which a class can be deleted are described below, as well as information about other classes with which this class is associated. Each property of the class is also listed in the following format:

Each method is described as follows.

Set Basically, a set is a component of multi-device management and contains objects that group devices according to some logical or physical grouping. There is a one-to-many relationship between the Sets object and the Devices object. The Execute method can execute a command on each member of the Sets object.

An instance of the Sets class is created by the user (or other process). In one embodiment, instances are never automatically created by other models. The only property needed for a new instance of the Sets class is Name. The value of Name needs to be unique among the instances of Sets on the controller. An instance of the Sets class can be deleted by the user (or other process). Instances are never automatically deleted by the object model in the described embodiment. When an instance is deleted, other instances (including instances that reference the deleted class) are not changed at all.

  An instance of Sets can have an association to an instance of Devices. An instance of Sets has an association to each instance of Devices that is a member of the set. Each instance of Sets can be associated with zero, one, or multiple instances of Devices. An instance of Sets can also have an association with an instance of JobInvocations. If the JobInvocations instance is a template (JobInstructions RootJobID property is zero), the association indicates that the job template represented by the JobInvocations instance is to be executed on the association set. If the JobInvocations instance is a history record (RootJobID is not zero), the association represents the fact that the job in JobInvocations was executed on the association set.

Set property Name

The value of this property can be changed using the Rename method. The value must be unique among the Sets instances on the system.
Description

This is a free text description of the set. This description can be NULL or empty and can be updated.

Unused Properties The Caption property, InstallDate property, and Status property are inherited from the parent class, but are currently unused in the current embodiment.

Method AddDevice
Use AddDevice to add controlled or uncontrolled devices to the set.

  Remove Device

  Rename

  Execute

This method causes the job to be executed for members of the set where the job is executed. If a job is successfully created, it will cause a JobInformations instance to be created for the job. The value of the RootJobID property of the new JobInvocations instance is returned as the return value of this method. Note that job execution is asynchronous, so the success of this method does not mean that the job itself will work on the agent.

  The CommandType argument specifies how the Command and Parameter fields are to be interpreted.

This is written to the Command property of the new JobInformations instance. The value of the Parameters argument field is silently ignored if CommandType is 3. The maximum length of the Description argument is 256 Unicode characters. This is stored in the Description property of the new JobInformations instance. This is not an error about the set being empty. In this case, the JobInformations instance is created as usual with the parent Jobs instance. However, there is no child Jobs instance. The user of the object model must be prepared for this situation.

If the special command CommandType is 3, the Command argument contains one of the following values:

The text is case insensitive (ie, “Reboot”, “REBOOT”, and “reboot” all cause the server to reboot). If the value of the Command argument is not one of the above, the method returns an error.

Devices
Devices is a member of the set. A set is a group of devices. Devices represents, for example, a physical computer system that is typically a server device. In multi-device management, one goal is to manage multiple machines simultaneously. Nevertheless, commands can also be executed on device objects.

Creation An instance of Devices can be created at the request of a user or other process, or can be created automatically by the controller based on receipt of an auto-detection packet from an agent. Instances of the Devices class can be created manually by the user or by other processes. The new instance requires at least a value for the Name property, and this value cannot exist as the Name of another instance of the Devices class on the system. The value of the Name property needs to be a name that can be resolved on the controller up to the IP address of the management interface on the device itself. This resolution is usually done using a DNS server.

  A new instance can be associated with one or more MAC addresses, and each MAC address can be associated with one or more IP addresses. This is indicated by associating a new instance of the Devices class with an instance of DeviceHWAddrs and associating DeviceHWAddrs with DeviceHWIPAddrs. However, to communicate with the machine, only the name is required. The controller uses DNS to resolve names to IP addresses. MAC and IP are not used for communication, but are basically used only as information on the controller. When the instance is created, the controller sends a request to the agent to obtain node information including the IP address and MAC address. For all Devices instances created manually, LastDiscoveryTime is NULL. If at some later time an auto-detect packet matching this Devices instance is received, this field is updated with the time the packet was received.

  Alternatively, the SMBIOS GUID can be stored in the controller and used to uniquely identify the device. This can be done without any changes to the object model. This is because the SMBIOS GUID can be stored as a new device type (in DeviceHWAddrs linked to DeviceTypes representing the SMBIOS GUID).

Delete An instance of the Devices class can be deleted by a user or other process. The instance is Controller. If the RefreshDeviceList () method is called, it can be deleted by the system. When an instance of Devices is deleted, any instance of DeviceHWAddrs that references the device is also deleted (this may also cause the deletion of linked DeviceHWIPAddrs).

Associations Instances of Devices can have associations to the following classes:

For Sets instance A Devices instance has an association to each instance of Set that it is a member of. Each instance of Devices can be associated with zero, one, or multiple instances of Sets.

• For instances of JobInvocations An instance of Devices has an association to JobInvocations. If the JobInvocations instance is a template (JobInvocations RootJobID is zero), the association indicates that the job template represented by the JobInvocations instance is to be executed on the associated device. If the JobInvocations instance is a history record (RootJobID is not zero), the association indicates that the job in JobInvocations has been executed on the associated device.

A Device instance for a Job instance has an association to a Job instance that represents the parent job of each job executed on that device.

A Device instance to a DeviceHWAddrs instance has an association to a DeviceHWAddrs instance for each hardware address on the device that the controller has knowledge of. In the current version, the only hardware address stored on the controller is the NIC card MAC address.

A Device instance for a single instance of DeviceTypes has an association to an instance of DeviceTypes that gives the device type. In the current version, only a single device type is supported, so every instance of Devices is associated with a single instance of DeviceTypes.

Property Name
This is inherited from the parent class.

The value of this property can be changed using the Rename method.

  Alive

If the device is not controlled, this property is always false. Otherwise, this property is set to true each time a heartbeat is received from the device when the device becomes controlled (if it is currently set to false). . The heartbeat period is a global setting for the controller. The agent must send a heartbeat packet every time the heartbeat period. If the controller does not receive a heartbeat from the device for 1.5 times the heartbeat period, it sets this property to false. The heartbeat period used is a period currently set on the controller. This period may be different from the device heartbeat period.

  Controlled

The value of this property is set by starting or ending control of the device using the Manage method.

  HeartbeatTime

This is NULL if the device is not controlled or becomes uncontrolled. Otherwise, when the device comes to be controlled, this is set to the time when the device came to be controlled. While the device is in control, this is set to the time when the latest heartbeat was received from the device.

  AlertStatus

  If the device is not controlled or is no longer controlled, it is always set to Unavailable. Otherwise, when the device becomes controlled, it is initialized to Status not Known. After a call to EnableAlerts () is made, this is set to the value returned from the agent that can be Enabled, Enabled, or Disabled.

  Type

  Description

  LastDiscoveryTime

Each time an auto-detect packet is received that matches this Device instance, the date is updated to the current date. This can be used to identify manually entered records that were not previously on the network.

  The following properties are inherited from the parent class but are currently unused. Caption, CreationClassName, InitialLoadInfo, InstallDate, LastLoadInfo, NameFormat PowerManagementCapitalCapabilities, PowerRightCapabilities, PowerRightCapabilities, PowerRatio

Method EnableAlerts

This call is synchronous. If the alarm is enabled on the device, the device returns the current alarm. If the alarm is disabled, the controller deletes the alarm details for this device.

  RecoverManagerDevice

This method can be called if the user thinks that the configuration on the device has been destroyed. The controller sends the following information to the device:

-Request for device information-Control request to manage this device-Current heartbeat configuration information (heartbeat interval)
Current alarm status for this device (enabled or not enabled)
This call returns after all the above has been processed and may therefore take some time.

  Manage

This changes the state of control for the device and returns when the change is complete. The value of ControlFlag specifies the operation to be performed on the device.

If an attempt to control the device fails due to an incorrect protocol version, Error is returned.

  Execute

The SetPowerState method is inherited from the parent class, but is currently unused in this embodiment.

DeviceTypes
In one embodiment, the device reports the same device type, “Microsoft Server Appliance”.

Instances of DeviceTypes can be created manually or automatically. When manually created, the Name needs to be unique among the entire existing DeviceTypes instance. Manual creation can be used on networks where automatic detection does not work.

  An instance of DeviceTypes can be created automatically based on incoming auto-discovery packets or based on the creation of an instance of Devices. Another way in which an instance of DevicesTypes is created is based on the creation of an instance of Devices. Each Devices instance includes a DeviceType string. If an instance of DeviceTypes corresponding to the above string does not exist, a new instance of DeviceTypes is created that has the above string as its Name and a Blank Description.

Deleted DeviceTypes instances can be deleted, but can only be deleted if there are no instances of Devices with the same type of string as the deleted instance. If deletion of this reason fails, the WMI DeleteInstance method must return with the error WBEM_E_FAILED.

Associated An instance of DeviceTypes is associated with each instance of the same type of Devices.

Property Name

The device type name cannot be changed.

  Description

  The Caption, InstallDate, and Status properties are inherited from the parent class, but are currently unused in the current embodiment. There is no method for this option.

HWADDRTYPES
All devices typically include a number of individually identifiable hardware parts. Each hardware part can contain an address, ie a unique identifier.

In one version, the only address type used internally is “MAC”. This is used to store the device's MAC address. An instance of HWAddrTypes can be created manually or automatically. An instance is automatically created if an instance of DeviceHWAddrs is created when the Type property value does not match the Name property of an existing HWAddrTypes instance. In this case, a new instance of HWAddrTypes is created with a Name property that has the same value as the value in the DeviceHWAddrs Type property, and the Description property is blank.

Delete An instance of HWAAddrTypes can be deleted. Deletion fails if any instance of Devices contains the same Type property as the Name property on the deleted instance.

The associated HWAddrTypes instance is associated with all instances of DeviceHWAddrs that contain hardware address information for this type of hardware.

Property Name

  Description

  The Caption, InstallDate, and Status properties are inherited from the parent class, but are currently unused in the current embodiment. There are no methods on this object.

DEVICEHWADDRS
The device hardware address uniquely identifies a hardware component such as a NIC.

In one current embodiment, the only hardware address type used internally is “MAC”. This is used to store the device's MAC address. An instance of DeviceHWAddrs can be created automatically or manually. The automatic creation is performed based on the received automatic detection packet. If the packet contains a hardware address (including type and address), a new instance of DeviceHWAddrs will be created for that hardware address (this may also involve creating an instance of HWAddrTypes) ).

Delete An instance of DeviceHWAddrs can be deleted. Also, the instance is automatically deleted when the corresponding Device instance is deleted.

Association For a single instance of Devices An instance of DeviceHWAddrs is associated with an instance of Devices that contains a hardware address.

For a single instance of HWAddrTypes An instance of DeviceHWAddrs is associated with an instance of HWAddrTypes that defines the type of hardware address stored in this DeviceHWAddrs instance.

For an instance of DeviceHWIPAddrs An instance of DeviceHWAddrs is associated with zero or several instances of DeviceHWIPAddrs for each IP address associated with this instance of DeviceHWAddrs. This association is only used for DeviceHWAAddrs that contain the address of the NIC hardware.

Property HWAAddr

  DeviceName

  Type

Unused Properties The Caption property, InstallDate property, Name, and Status properties are inherited from the parent class, but are currently unused in the current embodiment. There are no methods on this object.

DEVICEHWIPADDRS
This object is not currently used.

JOBINVOCATIONS
There are two categories of Job Innovations. The first category is a template. There are Job Invocations with a RootJobID equal to zero. That is, they are simply templates and are not related to a specific execution. The second category of Jobs Informations is history. That is, these are job call history records.

Instances of JobInvocations can be created manually or automatically. Job templates are created manually, while job history records are created automatically. Instances of JobInvocations can be created manually (or by other processes) by the user. Only a job template can be created (by definition, an instance of JobInvocations with a RootJobID of zero), and it is an error to attempt to create a JobInvocations instance with a non-zero RootJobID. An instance created manually by JobInformations can later change any of its writable properties.

Automatic creation (job history)
When a job is executed (using Devices.Execute or Sets.Execute), an instance of JobInvocations is created. This is given a new unique value (not zero) for RootJobID. The properties of the new instance, Name, Command, Parameters, and Description are the Devices. Execute method or Sets. Populated with the values of JobInvocationName, Command, Parameters, and Description arguments for the Execute method. The properties TargetName and TargetType are filled with the name of the set or device on which the job is running, and the type of set or device. The properties of automatically created JobInformations instances cannot be changed. Attempting to change the properties of a JobInvocations instance in WMI causes the PutInstance (Put_from script) method to return WBEM_E_FAILED (WbemErrFailed from scriptOnly) and the job template (RootJob ID and history job ID is zero) (If the RootJobID is not zero) can be deleted, and if the job history is deleted, all related Jobs instances and JobLogs instances are also deleted, which is implemented in WMI.

Related • For a single instance of Devices An instance of JobInvocations is an instance of Devices if the job template is defined to run on that device, or if the job history is run on that device. Is related to. This association exists only if the value of TargetType is Devices, and if so, the association is for a Devices instance that has the same name as given in the TargetName property.

For a single instance of Sets An instance of JobInvocations is set to an instance of Sets if a job template is defined to run on that set, or if job history is run on that set. Related. This association exists only if the value of TargetType is Sets, and if so, the association is for a Sets instance that has the same name as given in the TargetName property.

For a single instance of Job An instance of JobInstructions is associated with an instance of Job that gives the result of executing the job represented by JobInvocations. This association exists only when the JobInformations instance is a history record (ie, the RootJobID is not zero).

Property RootJobID

Root job unique identifier. This can be any 64-bit value. If this value is zero, it indicates that this JobInformations instance is a template, not a history record for a previously executed job. If this value is non-zero, this instance is a history record.

  Name

This is because of Devices. Execute or Sets. Stores the name of the job passed to Execute as a JobInvocationName argument.

  TargetName

The target name, Devices or Sets where the job was run
TargetType

  Command

The command that is called on the target.

  Parameters

Parameter Description for the command to be called

Unused properties The following properties are inherited from the parent class, but are currently unused. Caption, ElapseTime, InstallDate, Notify, Owner, Priority, StartTime, Status, TimeSubmitted, UntilTime.

Method Rename

JOBS
The Jobs object captures the Job Innovation topology. For example, if a user has a set object and invokes Execute on that set, there will be one Job Object created as a Root Job, and there will also be a Jobs object created for each device in the set. Exists. The Root Jobs object is used as an entry for the Jobs topology structure. In one embodiment, only parent-child relationships are used, but the model can be instantiated as a tree of depth N.

Jobs are automatically created by the system when a job is started (eg, by the method Sets.Execute or Devices.Execute).
An instance of Jobs cannot be updated. If a job instance representing a parent job is deleted, the following are also deleted:

The related JobInvocations instance The child instance of Jobs related to the Jobs instance The JobLogs instance related to each of the child Job instances This is the record of the job invocation, the status of the job on each device, and the result of the job on each device Is deleted.
It is possible to delete a job that is currently in progress. If this is done, no further output from the job received from the controller is stored at all and no errors are reported.

Association • For a single instance of JobInvocations An instance of Jobs is associated with an instance of JobInvocations that represents a job.

For an instance of JobLogs An instance of Jobs is associated with an instance of JobLogs that contains the output from the execution of this job on a particular device. This association exists only when an instance of Job represents a single device, not a parent instance.

For a job instance A job instance is associated with a parent job instance that represents a child process of the parent job instance. In the current version, only one level of parent-child relationships is supported, where the parent represents a job that runs on multiple devices and the children represent results from individual devices.

For a single instance of Devices An instance of Jobs is related to an instance of Devices if the Jobs instance holds the results of running a job on a single device. This association gives the device on which the job is running.

Property JobID

  ParentJobID

  DeviceName

  StartTime

  EndTime

JobStatus

Sets. Execute method or Devices. When the Execute method returns, the status value is one of 2, 3, or 11. If the job cannot be started on any device, the status is set to 2. If some of the devices cannot start the job, the status is set to 11. Otherwise, the job has been started on all devices and the status is set to 3.

  After the Execute method returns, the caller can poll the status value for the parent Jobs instance. The status is 3 or 11 if the job is still running on at least one device. If the job is completed on all devices, the status is 0 or 1. The status is 2 if no job has started on any device.

Valid values for child jobs are 0 and 3-10. The value is initialized to 3 when a child Jobs object is created. If an error occurs during job execution, JobStatus is updated to either 5 or 7-10. If the job is successfully transmitted to the device, JobStatus is set to 4.
Jobs. JobStatus is set to 6 when the job is stopped by the user calling Stop. If the job completes successfully, JobStatus is set to zero.

  The following properties are inherited from the parent class but are currently unused. Caption, Description, InstallData, Name, and Status.

Method Stop

This method stops a job that is running on the device. If this method is executed on the instance representing the parent job, all child chills still running are stopped. A job stopped by the user has a Status property value of “job stopped by user”.

  GetOutput

Values for OutputType are as follows:

• 0 = get exit status • 1 = get standard output • 2 = get standard error • 3 = get all outputs (sequentially)
This method is valid only for child Job instances. If OutputType is 0, the exit status is returned in the Output string. For example, 32 exit statuses are returned in Output as the string “32”.

JOBLOGS
JobLogs captures the output of script or execution file execution. The job log is related to Jobs. There can be N JobLogs for any Jobs.

JobLogs are always automatically created by the system when a job is started (eg, by method Sets.Execute or method Devices.Execute).
An instance of the JobLogs class cannot be updated, and an instance of the JobLogs class cannot be deleted in WMI. The association is for a single instance of Jobs. An instance of JobLogs is associated with the instance of Job that gives the device and job that generated this output.

Property JobID

  Sequence

Start at 1 and climb. The same sequence number used for stdout, stderr, exit status.

  LogTime

  OutputType

  OutputData

These properties are inherited from the parent class, but are currently unused. Caption, Description, InstallDate, Name, and status. There are no methods on this object.

ALERTS
The device hardware address uniquely identifies a hardware component such as a NIC.

The alarm object holds a copy of all alarms from managed devices that report alarms to the controller. Whether or not the device reports an alarm is determined by Devices. Set using the EnableAlerts () method. Each instance of the Alerts object represents a single alert from a single device. Alert instances are always automatically created based on information provided by the device. Alert instances cannot be created manually (in WMI).
Instances of this class cannot be updated, and deleting an instance of the Alerts class will cause the alarm to disappear on the device itself.
Associations are for a single instance of Devices, for a Device instance that represents the device on which the alert was created.

Property DeviceName

This includes the name of the device for which the alert was created.

  Cookie

  AlertType

  AlertID

  AlertLog

  AlertSource

  AlertReplaceString

This is stored in the format received from the device. Currently, this is a format as a list of values separated by newline characters (character code decimal 10). If the value contains a comma, no escaping is done.

  ReceivedTime

This is the time at which the controller received an alarm notification from the device.

  The following properties are inherited from the parent class but are currently unused. Caption, Description, InstallDate, Name, and Status. There are no methods on this object.

CONTROLLER
The Controller class is used to configure and control the controller. The controller has some global configurable parameters such as heartbeat interval. In addition, controller services and sub-services can be started and stopped.

This is a class with only a single element for creation and deletion, and this instance cannot be updated. There is no association at all.

Property HeartbeatInterval

If not specified, the value 120 seconds (2 minutes) is used. The minimum value that can be set is 60 seconds. The following properties are inherited from the parent class but are currently unused. Caption, Description, and Name.

Method RefreshDeviceList

The refresh device list is a process having two steps. First, uncontrolled devices are removed from the data store. Next, the claimed detection is initiated for a device having the “Microsoft Server Appliance” type. Then the method returns. At this point, the result of initiating detection may not have been received by the controller, so the Devices table may be empty or partially complete.

  SetHeartbeatInterval

This method is called Controller. Update the heartbeat interval in the HeartbeatInterval property. The method then sends this updated interval to all managed devices. Then the method returns. The method does not wait for the status of sending an updated heartbeat to the device. This value is given in seconds. The minimum value that can be set is 60. If an attempt is made to set the value below 60, the stored value is set to 60 in one embodiment.

If this value is increased, the device may appear to be inactive after 1.5 ^* old heartbeat interval (Devices.Alive property is set to false). Within WMI, special classes called association classes are used to link object instances. This section defines the related classes used to link WMI classes that implement the object model.
The related class is derived from CIM_Component or CIM_Dependency.

  CIM_Component includes a property GroupComponent and a property PartComponent that specify a parent instance and a child instance, respectively. CIM_Dependency includes a property Antecentent and a property Dependent that specify a dependency relationship. As described below, the parent class is listed and then the value of the property (GroupComponent and PartComponent, or Antecedent and Dependent) is described. None of the classes add additional properties or revoke descriptions or other attributes.

  DEVICEHWADDRTODEVICEHWIPADDR

GroupComponent is a reference to an instance of DeviceHWAddrs, and PartComponent is a reference to an instance of DeviceHWIPAddrs.

  DEVICEHWADDRTOHWADDRTYPE

  Antecedent is a reference to an instance of DeviceHWAddrs, and Dependent is a reference to an instance of HWAAddrTypes.

  DEVICEOALERT

  Antecedent is a reference to an instance of Devices, and Dependent is a reference to an instance of Alerts.

  DEVICEICE DEVICEHWADDR

GroupComponent is a reference to an instance of Devices, and PartComponent is a reference to an instance of DeviceHWAddrs.

  DEVICEICEDEVICETYPE

  Antecedent is a reference to an instance of Devices, and Dependent is a reference to an instance of DeviceTypes.

  DEVICETECOBINVOCATION

  Antecedent is a reference to an instance of Devices, and Dependent is a reference to an instance of JobInvocations.

  SETTOJOBINVOCATION

  Antecedent is a reference to the instance of Sets, and Dependent is a reference to the instance of JobInvocations.

  DEVICETECJOB

  Antecedent is a reference to an instance of Devices, and Dependent is a reference to an instance of Jobs.

  JOBTOJOB

  GroupComponent is a reference to an instance of Job (representing a parent), and PartComponent is a reference to an instance of Job (related to a child).

  JOBINVOCATIONTOJOB

  Antecedent is a reference to an instance of JobInvocations, and Dependent is a reference to an instance of Jobs.

  JOBTOJOBLOG

  Antecedent is a reference to an instance of Jobs, and Dependent is a reference to an instance of JobLogs.

  SETTODEVICE

  GroupComponent is a reference to an instance of Sets, and PartComponent is a reference to an instance of Devices.

100 Computer System Environment 110 Computer Device 120 Processing Device 121 System Bus 130 System Memory 131 Read-Only Memory 132 Random Access Memory 133 Basic Input / Output System 134 Operating System 135 Application Program 136 Other Program Modules 137 Program Data 140 Non-Removable Nonvolatile Removable memory interface 141 hard disk drive 144 operating system 145 application program 146 other program modules 147 program data 150 removable non-volatile memory interface 151 magnetic disk drive 152 removable non-volatile magnetic disk 155 optical disk drive 156 removable Non-volatile optical disc 160 User input interface 161 Mouse 162 Keyboard 163 Microphone 164 Tablet 170 Network interface 171 Local area network 172 Modem 173 Wide area network 180 Remote computer 181 Memory device 185 Remote application program 190 Video interface 191 Monitor 194 Output peripheral interface 195 Speaker 196 Printer 200 Architecture 202 Controller computer 204 ₁ , 204 ₂ , 204 _n- node computer 206 ₁ , 206 ₂ , 206 _n- node program 208 Controller service 210 Management program 212 Transport layer 214 ₁ , 214 ₂ , 214 _n Transport Layer 216 Physical Network 218 ₁ , 218 ₂ , 218 _n Agent Service 220 ₁ , 220 ₂ , 220 _n Node Operating System 222 Operating System 302 Data Store 304 Start Process 306 Controller Program 310 Schema Interface 312 Agent Program 314 XML Message Exchange Protocol 316 Transport Protocol 320 Script / Execution Engine 322 Reboot / Shutdown / Suspend Component 330 Discovery Component (Discovery Component)
332 detection listening component (detection listening component)

Claims (30)

A method for controlling a plurality of computer systems from a controller in a data center computer network, comprising:
Maintaining at least one unique set in the controller, each distinct set being distinguished by operational capabilities, customer relevance, or executing application and configured to process jobs Providing a group of devices;
Determining that a computing device belonging to at least one of the unique sets has completed part of the job assigned to the computing device;
Providing at the controller a new job corresponding to at least one operation performed on a computing device belonging to the plurality of unique sets;
Receiving a command to redeploy the computing device that has completed part of the assigned job, the redeploying comprising assigning the computing device to a new unique set;
Sending a message from the controller to each computer device in the new unique set comprising the redeployed computer device, wherein the message is sent to each computer device receiving the message Instructing the job to run,
Storing the result of the new job from each computing device in the new unique set at the controller.   The method of claim 1, wherein providing the unique set at the controller comprises providing data corresponding to at least one set of computing devices.   The method of claim 1, wherein providing the job at the controller comprises providing data corresponding to a script executed on the unique set.   The method of claim 1, wherein providing the job at the controller comprises providing data corresponding to a binary program executed on the unique set.   5. The method of claim 4, wherein the data corresponding to a binary program executed on the unique set comprises a network address.   The method of claim 1, further comprising receiving the message at an agent on a computing device identified in the unique set and executing the job in response to the message.   The method of claim 6, wherein executing the job in response to the message comprises executing a script.   The method of claim 6, wherein executing the job in response to the message comprises executing a binary program.   The method of claim 8, wherein executing a binary program comprises retrieving the program based on a network address in the message.   The method of claim 1, further comprising receiving discovery information at the controller indicating that a node computing device is operable to be controlled by the controller.   The method of claim 10, further comprising recognizing that the node computer device is already controlled by the controller.   The method of claim 10, further comprising recognizing that the node computer device is not controlled by the controller and controlling the node computer device.   The method of claim 12, further comprising adding information identifying the node computer device to a data store maintained by the controller.   The method of claim 10, further comprising automatically configuring the node computer device based on receiving the discovery information.   The method of claim 1, wherein storing the results of the job comprises collecting the results in a storage.   The method of claim 1, wherein storing the result of the job comprises persisting the result. A system for controlling a plurality of computer systems from a controller in a computer network of a data center,
A controller configured to hold a plurality of unique sets, each set comprising a group of at least one computing device configured to process a job, each unique set being capable of operation, customer Controllers that are related to or running applications,
A computer device identified in the unique set comprising agent software connected for communication with controller software on the controller;
A job held by the controller corresponding to at least one operation performed on the computing device identified in the unique set, wherein the computing device processes the job and then processes the new job A job configured to, and
After determining that the computer device has finished part of the job assigned to the computer device, a message including data corresponding to the job is communicated from the controller software to the agent software of the computer device. Wherein the message causes the agent software to make part of the assigned job part of a new unique set for the computing device that has finished and to execute the new job, A transport in which the agent software of the computing device executes the new job in response to receiving the message and returns a result to the controller;
A data store of the controller, wherein the controller comprises a data store for storing the result from the agent software in the data store.   The system of claim 17, further comprising a schema interface configured to provide access to information in the data store.   The system of claim 17, further comprising an execution engine in the node computer, wherein the agent software communicates with the execution engine to perform at least one operation corresponding to the job.   The system of claim 19, wherein the execution engine comprises a script engine, and the agent software communicates with the execution engine to execute a script.   The system of claim 19, wherein the execution engine software executes a binary program, and the agent software communicates with the execution engine to execute the binary program.   18. The system of claim 17, further comprising software on the computing device that performs the set of at least one special operation requested by the controller.   The system of claim 22, wherein the set of at least one special operation comprises a reboot operation.   The system of claim 22, wherein the set of at least one special operation comprises a suspend operation.   The system of claim 22, wherein the set of at least one special operation comprises a shutdown operation.   The system of claim 17, further comprising a discovery listening process in the controller that detects discovery information provided by a computing device on the network.   27. The system of claim 26, wherein the controller includes software for automatically configuring a computing device that provides the discovery information.   27. The system of claim 26, wherein each computing device includes a discovery component for automatically providing the discovery information.   29. The system of claim 28, wherein each computing device automatically provides the discovery information following a reboot of the computing device. A method for controlling a plurality of computer systems from a controller in a data center computer network, comprising:
Maintaining multiple unique sets in the controller, each unique set being identifiable by network address, distinguished by function of operation, customer relationship, or executing application, processing binary programs Comprising a group of at least one computer device configured to:
Determining that a computing device belonging to at least one of the unique set has finished processing a portion of the binary program assigned to the computing device;
Providing at the controller a new binary program comprising at least one script for processing on a computer device belonging to the plurality of unique sets;
Receiving by the agent a command to redeploy the computer device that has completed part of the assigned binary program;
Redeploying the computing device that has terminated a portion of the assigned binary program in response to the command, wherein the redeploying assigns the computing device to a new unique set. Preparing,
Sending a message from the controller to each computer device in the new unique set comprising the redeployed computer device, wherein the message is sent to each computer device receiving the message Instructing to execute a binary program;
Storing the result of the new binary program from each computer device in the new unique set at the controller.

Images